Beyond aerobic intensity: Comparing the effects of cognitively engaging gross-motor coordination training and aerobic...

Working memory and selective attention develop rapidly during middle childhood and are sensitive to physical activity. Aerobic exercise is thought to act mainly through intensity-driven physiological pathways (cerebral perfusion, neurotrophic signalling and arousal), whereas gross-motor coordination training additionally imposes perceptual, decision-making, motor-planning and inhibitory demands that may provide stronger and more domain-specific cognitive stimulation. This study compared 14-week gross-motor coordination and aerobic exercise interventions, designed a priori to be matched for session frequency, duration and rated perceived exertion, on working memory and selective attention (indexed by processing volume on the d2 test) in children aged 10-12 years.

The study employed a pre-test-post-test randomized controlled design. Participants (n = 78) were recruited from a single state primary school and allocated by computer-generated stratified block randomisation (1:1:1) to an Aerobic Exercise Group (AEG; n = 26), a Coordination Exercise Group (CEG; n = 26), or a Control Group (CG; n = 26). Intervention groups participated in structured exercise programmes consisting of 40-minute sessions three days per week for 14 weeks, delivered on school premises during the school day as an addition to the standard physical education curriculum. Working memory was assessed using the examiner-administered, paper-and-pencil Visual-Auditory Digit Span Task, and selective attention was measured with the paper-based d2 Test of Attention; the digit span was administered individually and the d2 in small groups, both in a quiet classroom and by assessors blinded to group allocation. No computerised platform was used. Data were analysed with aligned rank transform ANOVA (ART-ANOVA).

The group × time interaction was statistically significant for both outcomes (working memory: F(2, 75) = 12.47, p < 0.001, partial η² = 0.25, 95% CI [0.09, 0.39]; selective attention: F(2, 75) = 8.23, p < 0.001, partial η² = 0.18, 95% CI [0.04, 0.32]). Both intervention groups demonstrated significant improvements in working memory and selective attention compared with the control group (all Bonferroni-adjusted p ≤ 0.018). When the intervention groups were compared directly, the coordination-based exercise programme was significantly more effective than the aerobic exercise programme in improving working memory (adjusted p = 0.045; median change 5.2 [IQR 4.6-5.9] vs. 3.1 [2.2-3.9] points on the digit span total score; between-group effect size r = 0.31, 95% BCa CI [0.06, 0.52]; median difference 2.1 points, 95% BCa CI [1.44, 2.76]). No significant difference was found between the two intervention groups with respect to selective attention (adjusted p = 0.320; median change 29.0 [17-42] vs. 30.5 [25.0-40.0] d2 TN points; r = 0.16, 95% BCa CI [- 0.10, 0.40]; median difference - 1.5 points, 95% BCa CI [- 10.48, 7.48]). The correlation between working memory and selective attention change scores was weak and statistically non-significant (Spearman's rho = 0.12, p > 0.05).

Both aerobic and coordination-based exercise interventions may support working memory and selective attention in children. Coordination-based interventions that were designed a priori to impose a higher cognitive load - through motor planning, attentional switching, inhibition of prepotent movement responses, and progression from closed to open skills - may yield stronger gains in working memory specifically, although the between-intervention difference was modest and requires replication. Because selective attention was indexed by the number of items processed, the attentional findings should be read as evidence about processing volume rather than about attentional efficiency in the narrow sense. School-based physical activity programmes may therefore benefit from being structured not only around movement intensity but also around cognitive engagement and coordination components. Registration was not completed before the first participant was enrolled, which the authors acknowledge as a departure from best practice. No changes were made to the eligibility criteria, outcomes, or planned analyses after enrolment began; the primary outcomes, group allocation and analysis plan reported here are identical to those documented in the ethics application dated 27 February 2026, which is available on request. The authors have confirmed that this journal permits retrospective registration provided it is declared.

ClinicalTrials.gov, NCT07658404, registered 13 June 2026. Retrospectively registered.